graphlit's profile picture. Context engineering for developers. Ingest any content, extract what matters, give your AI the memory it needs.

Graphlit

@graphlit

Context engineering for developers. Ingest any content, extract what matters, give your AI the memory it needs.

Pinned

🔥 New @graphlit platform release! In addition to our MCP server, we've released support for OpenAI GPT-4.5, Sonnet 3.7, and the new @MistralAI OCR model for document preparation. changelog.graphlit.dev


Graphlit reposted

It’s just a commodity RAG service. Has tons of limitations which any other RAG API supercedes. For the simple stuff, it’s great if you want to stay in the Google ecosystem. But for real production content, you need a platform like @graphlit or one of the other vendors.


Graphlit reposted

Memory Is the Interface. Knowledge Is the Substance. 🧵1/n


Graphlit reposted

It’s been interesting to see so many companies in the agents and memory space optimizing for onboarding. “In five lines of code” “Integrate in 5 min” While that’s important to avoid friction, it’s not a concern we hear a lot at @graphlit The questions we get are more…


Graphlit reposted

💡We've rewritten our @graphlit docs from scratch: docs.graphlit.dev And we have refreshed the AskGraphlit app (ask.graphlit.dev) and embedded it in our developer portal. 🆕Even more exciting, we now have an MCP server to use with Claude Code, Cursor, etc. Add…


One week to build agentic workflows. Who needs help?

Who is looking for: - data connectors (Gmail, Notion, Jira, etc) - automatic RAG-ready ingestion - hybrid + metadata retrieval - MCP tools What can we build for you in a week? We’ve helped startups go from 0-1 in days (including weekends). Look at @zine_ai for ideas -…



Graphlit reposted

Hybrid similarity search... from a voice note? Yes w/ @zine_ai Here I'm capturing a short audio clip, which is transcribed live with @AssemblyAI. From that 'find similar' uses the transcript to locate other similar contents, and the top hit is the summary of the PDF I was…


Graphlit reposted

Seamless filtering of your LLM context w/ @zine_ai Automatic 'observations' as a knowledge graph. Filter by any label, person, organization, etc. Can even 'find similar' by text or image. Build context, chat with any LLM, or share via MCP.

KirkMarple's tweet image. Seamless filtering of your LLM context w/ @zine_ai 

Automatic 'observations' as a knowledge graph.

Filter by any label, person, organization, etc.

Can even 'find similar' by text or image.

Build context, chat with any LLM, or share via MCP.
KirkMarple's tweet image. Seamless filtering of your LLM context w/ @zine_ai 

Automatic 'observations' as a knowledge graph.

Filter by any label, person, organization, etc.

Can even 'find similar' by text or image.

Build context, chat with any LLM, or share via MCP.

Graphlit reposted

Sneak peek... ingest anything into Zine. Automatic transcription and chapters for audio/video. Ask questions about the content (supports images too). Way more to come... stay tuned. Powered by @graphlit.

zine_ai's tweet image. Sneak peek...  ingest anything into Zine.

Automatic transcription and chapters for audio/video.

Ask questions about the content (supports images too).

Way more to come... stay tuned.

Powered by @graphlit.
zine_ai's tweet image. Sneak peek...  ingest anything into Zine.

Automatic transcription and chapters for audio/video.

Ask questions about the content (supports images too).

Way more to come... stay tuned.

Powered by @graphlit.

Graphlit reposted

✨ Brand-new DX for agentic tool calling in @graphlit We've integrated LLM clients natively into a single 'streamAgent' method, which takes the tool definitions, tool callback methods, and the model specification. Just assign your OpenAI, Anthropic or Google API keys, and we…

KirkMarple's tweet image. ✨ Brand-new DX for agentic tool calling in @graphlit 

We've integrated LLM clients natively into a single 'streamAgent' method, which takes the tool definitions, tool callback methods, and the model specification.

Just assign your OpenAI, Anthropic or Google API keys, and we…
KirkMarple's tweet image. ✨ Brand-new DX for agentic tool calling in @graphlit 

We've integrated LLM clients natively into a single 'streamAgent' method, which takes the tool definitions, tool callback methods, and the model specification.

Just assign your OpenAI, Anthropic or Google API keys, and we…

Graphlit reposted

Context is everything. Call it RAG, RAG-as-a-service, knowledge base, memory, whatever you want… But now we are seeing everyone focus on providing context to LLMs. @cohere @MistralAI now have agent SDKs with builtin tools for context retrieval. IDEs like @cline and…


Graphlit reposted

So, how much are you spending per day on LLM tokens inside @cline, @windsurf_ai , etc?


Read more about the history of @graphlit - and how we're bringing back our old Unstruk Lens app. With the coding power of @anthropic Claude Sonnet 4 and using AI IDEs like @cline, we'll track how long it takes to bring this app back to life with Next.js and @vercel .

Little history about @graphlit - we started the company 4 years ago to tackle the problem of ingesting and organizing unstructured data: especially visual and geospatial data. This was in the pre-ChatGPT days, but we were using computer vision models and some early text analysis…



Graphlit reposted

Good to see @graphlit MCP in the wild!

Rocking codex with codex-mini-latest and MCPs :)

jasonkneen's tweet image. Rocking codex with codex-mini-latest and MCPs :)


Graphlit reposted

Our MCP search engine (McPoogle) now has an MCP server that works with @cline. If you're looking for an example of using the @vercel MCP Adapter to build an MCP server, check this out: github.com/graphlit/mcpoo…


Graphlit reposted

Memory is the hot button term these days. But offering “canonical knowledge” is something we’ve been doing at @graphlit for several years. Auto-organizing unstructured data of multiple formats; making it searchable and useful via summarization, extraction or publishing.…

What happens when AI has a better memory than your exec team? When AI can hold your company's entire knowledge base in working memory, the productivity leap becomes exponential. Context windows expanding from 100K to 10M+ tokens is the difference between an AI system that can…



Graphlit reposted

✨ Excited to announce the🥇MCP search engine: McPoogle (aka MCP-oogle). 🦝 Hosted by Inspector McPoogle, our raccoon mascot. Built on the @graphlit platform, you can search and ask questions about 4000+ MCP servers and tools. Here we're asking about MCP servers to use with…

KirkMarple's tweet image. ✨ Excited to announce the🥇MCP search engine: McPoogle (aka MCP-oogle).

🦝 Hosted by Inspector McPoogle, our raccoon mascot.

Built on the @graphlit platform, you can search and ask questions about 4000+ MCP servers and tools.

Here we're asking about MCP servers to use with…

Graphlit reposted

Claude 🤝 @graphlit MCP Server Now you can create images right in Claude with the OpenAI image gen model.

KirkMarple's tweet image. Claude 🤝 @graphlit MCP Server

Now you can create images right in Claude with the OpenAI image gen model.
KirkMarple's tweet image. Claude 🤝 @graphlit MCP Server

Now you can create images right in Claude with the OpenAI image gen model.

Loading...

Something went wrong.


Something went wrong.